Depression in the elderly is complicated by both cerebrovascular disease and neurodegeneration, both of which are believed to contribute to the limited efficacy of antidepressant treatment in the elderly. However, the neurobiologic basis of the depressive syndrome and treatment response in late-life depression have not been established. The primary aim of this R01 application is to characterize the functional neuroanatomy of geriatric major depression, which we believe is characterized by altered functional connectivity, and use this to explain treatment response variability. The proposed study will identify in elderly individuals the changes in regional brain activity associated with being depressed, being treated for depression, and responding to depression treatment. To this end we will investigate, with fMRI, eighty elderly depressed subjects and 40 elderly controls. Subjects will undergo fMRI scanning on two occasions 12 weeks apart. For the depressed subjects, the scanning will occur just before and 12 weeks after initiating treatment with escitalopram. The depressed subjects will be restricted to individuals aged 65 and older, whose first episode of depression started at age 60 or older. We limit our sample to these 'late-onset'elderly depression subjects because, by definition, these individuals did not have mid-life depression, and thus should be most likely to show late-life specific biological factors. Our fMRI tasks target three of the key cognitive and affective neural pathways associated with LLD: a) cognitive control, b) declarative memory, and c) affective reactivity. These are central to theories of LLD and are associated with specific brain regions that have been linked to the neurobiology of LLD: the lateral prefrontal cortex (LPFC), the dorsal anterior cingulate cortex (dACC);the medial temporal lobe (MTL), and the amygdala. In the specific tasks, subjects will inhibit a prepotent response (cognitive control), recognize previously seen words (declarative memory), and respond to faces expressing emotion (affective reactivity). Functional connectivity, as well regional activity, will be estimated using the BOLD fMRI signal. The results of this study will characterize the functional neuroanatomy of late- life depression, which will be used to explain why some patients do not respond well to anti-depressant treatment. These treatment response subgroups can then serve as targets for future prevention and treatment studies.